how much faster is CCE vs tmpgenc

[jzchief]

I have been using TMPEnc for about a year now to create SVCD's from DivX or Xvid. (happy decent results)

I now have a DVD +-RW and have begun research on which path to take for converting DivX or Xvid to DVD.

I have been reading that quality is probably better with TMPGE but CCE is faster. I realize this will vary based on the system but how much faster. this may sway me.

I happen to be using:

PII 1GHz
256ram
80 gig hard drive
win 2000
TMPGE v 2.56.39
Vdub1.4.1
Nero 5.5.9

Thanks

ps can I reuse the TMPGEnc or do I need to upgrade.?

[jaxxboss]

PII ? WOW. Software isnt going to fix your speed woes.
Upgrade to a commodore 64

[q1aqza]

Much faster - on average usually 3-5 times depending on what you are doing.

[spiderman2k1]

I think he mean's P3. There is no 1gig p2 the slowest P2 was 233 and the fastest p2 was 450.

It's very fast not as fast as a transcode but the fastest encoder there is. You might save four hours in encoding time.

[jzchief]

You are correct P3. at 1G not P2

[FulciLives]

I started using CCE instead of TMPGEnc and have found that CCE is MUCH faster and the quality of the output is just as good from what I can see (I really wouldn't rate it better or worse but about the same).

Note I use CCE for doing DVD encoding ... I have heard that TMPGEnc can look better for VCD/SVCD whereas CCE can look better doing DVD.

As I said both look right about the same at DVD but I haven't done VCD/SVCD in a long time and when I did I only used TMPGEnc so I cannot comment on the "fact" that CCE is not as good in this regard.

However please realize that if you will be using CCE you better either already be good at using AviSynth AVS scripts or make peace with the fact that you will have to learn. In short TMPGEnc is much easier to use.

However the speed of CCE over TMPGEnc is well worth the learning curve.

- John "FulciLives" Coleman

[lordsmurf]

I don't see much speed increase at all, maybe 10% at most, given that both programs are set to the same quality levels and set for optimum performance settings.

CCE adds noise to video that annoys me.

[adam]

On average CCE is about twice as fast as TMPGenc. If you are encoding from a DVD source than this margin increases substantially since you can skip the conversion to RGB with CCE, but with your divx and xvid sources this is a non-issue. On multiprocessor setups TMPGenc can compete more in the speed department, but otherwise its one of the slowest, albeit highest quality, software encoders on the market.

Both encoders come with trial versions so just test them out yourself.

[Gazorgan]

CCE is definately 'faster'. However, most people to 3-4 pass VBR to compare to TMPGEncs 2-pass VBR and all speed advantages vanish.

Personally I've had great success doing CQ mode in TMPG. Speed is much faster than 2-pass VBR and quailty is as good if not better (it's a subjective thing).

[Vejita-sama]

Actually I can perform a 3pass VBR encode in CCE faster than a CBR encode in TMPGenc. In my hands CCE is ~4x faster than TMPGenc, and does a 3pass VBR takes about as along as a CBR encode.

As for quaility? I think that CCE results in better MPEG2 encodes than TMPGenc (total encode counting multipass VBR). At high bit rates either does the job. For MPEG1 encodes I think TMPGenc does a slightly better job (I forget the max supported bitrate for MPEG1 but it's fairly low).

In the end thou quaility is a matter of opinion. Just try out both programs demos and decide for youself.

[cooljoe666]

I just recently switched to CCE after a long haul with TmpgEnc and I definitely agree that CCE is much faster. TmpgEnc has too many problems with converting audio and video together and I believe it is better to do the two seperately and mux them later and that makes both programs very good but with CCE still having the edge. meow.

[Ripper2860]

I have to give the nod to CCE because of the speed and negligible quality differential (IMHO).

If you want the biggest quality/speed bang-for-the-buck, try CCE in conjunction with DVD2SVCD and D2SRoba. 3pass VBR quality in OPV (one-pass VBR) time frames.

If you are a DVD2SVCD and CCE user, you owe it to yourself to try it! NOS for your encodes !!

http://home.tiscali.no/tylohome/

** Can you tell that I'm a convert !!

[Kdiddy]

It's amazing how you drop off the board for close to 2 years, come back, and see the EXACT same topics being discussed. Just food for thought, carry on.

[andie41]

Kdiddy, I was sort of thinking along the same lines. When I first started following this board a couple of years ago, I was quite concerned because I could not get CCE to outperform TMPGenc at all, let alone by a factor of 4 - still can't. I can't even get Mainconcept to outperform TMPGenc, when all settings are virtually identical and I get a similar size and quality Mpeg. Oh well, such is life ...

[Ripper2860]

No disrespect, but this is the NEWBIE Conversion forum and as people enter the "black art" of video conversion/encoding, I am sure the same questions get asked over and over by different folks. I would think if any forum would be bring a sense of Deja-Vu, it would be the Newbie forums.

While I do have to agree that the difference in speed between TMPGenc and CCE is minor (certainly not 4x) if any (when configured properly), the real difference maker for me is the fact the D2SRoba3 only supports CCE with it's "One Pass VBR" and yields quality indistinguishable from TMPGenc 2-pass VBR in considerably less time than a 2-pass VBR run. Without the D2SRoba3 factor, it is a coin toss as to which is quicker -- again, assuming porper settings. Now throw in a P4 with SSE2 and I would think TMPGenc would take the lead.

As for which of the 2 encoder offers the greatest tools and flexibility -- TMPGenc, Hands Down!!

[adam]

Those people who argue that CCE isn't significantly faster than TMPGenc are simply performing flawed comparisons, period.

On a multi-processor system, yes CCE isn't much faster than TMPGenc. Other than that, it is between 2 and 4 times faster.

Remember the first pass in CCE is the creation of the .vaf file. So if you set it to 2-pass VBR you are actually doing 3-pass VBR. So a "2-pass" encode in CCE compared to a "2-pass" encode in TMPGenc is NOT a proper test. Learn how each encoder works, then make a comparison.

I think by far the biggest flaw in people's tests are that they fail to consider the fact that they just have a bottleneck which exists outside of the encoder. Unless you are using a raw avi souce, the speed at which you encode can only be as fast as the speed at which you decode your source. Even on a slow pc CCE can achieve double time per pass, but some codecs cannot even be decoded that fast. The same effect can be achieved by using certain filters or excessive filters. The limiting factor here is the speed at which the data is being fed into the encoder, NOT the speed of the encoder. With full D1 material I get double time encoding in CCE and about .8 in TMPGenc, but if I run a simple inverse telecine script I get about .65 in both encoders. Obviously this encode tells you nothing about the speed of the encoders, only of the speed of the filter.

Also don't forget that one of CCE's greatest advantages is that it can accept YUV sources natively whereas most other encoders have to convert it to RGB just to convert it back to YUV again. Given a YUV source and proper handling of it, ie: avisynth, this increases encoding speed by up to a whopping 40%.

The fact that CCE is significantly faster (2-4x) than TMPGenc has just been tested and retested on so many different levels than it amazes me some people still don't believe it. The author of TMPGenc himself has admitted it several times in interviews. This is not up for debate and its certainly not a coin toss. I love TMPGenc and it is a wonderful encoder but raw encoding speed just isn't one of its selling points.

[Ripper2860]

My apologies for propogating what appears to be a lie.

Me scampers off with my tail between my legs, looking for a place to hide and lick my wounds...

[andie41]

I had promised myself that I was finished with trying and testing encoders. I found something that works for me and planned to leave well enough alone. BUT, I am intrigued and if I can reduce my encoding time for a 60 minute video from approximately 130 minutes down to approximately a half hour, for around $60.00, then I am interested.

Did I understand this correctly?

Also, I would really like to read some of the interviews with Mr. Hiroyuki Hori (is that the author of TMPGenc?)- that would be very interesting. Also, Adam, if you could point me to some of the tests indicating its speed, I would be most appreciative.

[offline]

(Yeah, old argument but a good one)

I have to agree with Adam. CEE is the fastest but TMPGENC is not
too far behind on a p4 with HT using optimised filters and
frame serving with avisynth 2x. I do not have raw results
to show - sorry.

Canopus Procoder should not be overlooked, especially if you
are dealing with DV as AVI type 2 on a p4 with HT- it is amazing!
Other than that, I'll stick to CCE for mpeg-2 and TMPGENC for
general mpeg-1 and mpeg-2 from VHS/analog sources.

[SatStorm]

CCE IMHO is about twice faster TMPGenc Plus

What makes the huge difference between the two, is the use of avisynth. I don't know anyone using avisynth with TMPGenc, but I know almost anyone use it with CCE.

What Avisynth makes, is that frameserves really fast, resize the best possible way and filter fast.

Frameserving is neccessary, so not to have issues from your source to your encoder. Mpeg 4 users could benefit the most from this.
Resizing is an issue which only few really realise. Do a proper resizing and say goodbye to the fade in/outs blockness, those line antializing ungle artifacts, those blocks appeared on any static picture -expecially the dark ones-, etc. Which, by the way, are the "flaws" anyone "see" on TMPGenc and noone "see" on CCE.
Filtering is a must. Avisynth do it faster and if the source is from any mpeg, better. If the source is from any analogue capture, avisynth do it faster but the same with virtualdub. After all, those 2 programs use the same filters...

IMHO, after a minimum bitrate point and beyond, a value depending high on the final framesize, it is not the multipass VBR which make the difference. For example, you can see the difference with a 2 pass VBR vs a 5 pass VBR on a 352 x 576 framesize @ 2000kb/s, but on 2 Pass VBR Vs 5 Pass VBR on 352 x 576 framesize @ 3000Kb/s gonna look indentical IMHO. Those other things (resizing, filtering) are the ones who make the quality difference.

The ones who use TMPGenc, simply want to load any source to it, load a template and hit encode. No frameserving at all, and there are plenty users who also add filters from the built in ones. Well, you can't compare this with the cobo of CCE/avisynth...

The only other CCE benefit I know, is the YUV2 colourspace. It can handle it. From the other hand, TMPGenc has to convert anything to RGB (because of the use of the vfapi plug in). This is a quality factor for the NTSC users and not so a huge issue for us, the PAL ones.

My opinion, is that if you use avisynth with both encoders, CCE has better speed (about twice TMPEG's) and a better colour handling, which for the NTSC users is more than a must issue. But also has more "mosquito" noise, expecially from analogue sources. So, from DVD/DVB sources, CCE do a better job, from analogue sources, TMPGenc IMHO is better.
But I repait: You have to use with both encoders avisynth to test fair!

[Kdiddy]

No disrespect, but this is the NEWBIE Conversion forum and as people enter the "black art" of video conversion/encoding, I am sure the same questions get asked over and over by different folks. I would think if any forum would be bring a sense of Deja-Vu, it would be the Newbie forums.

None taken, but simply that is what the SEARCH option on this board is for.

[GeoPappas]

If CCE is so good, then why doesn't it have any ratings/votes/reviews in the Tools section?

When I search for tools, I usually sort them by their rating. I figure that highly rated apps are the best to try out. Since CCE didn't have any ratings, I never thought to try it.

[p0rnLiME]

Those people who argue that CCE isn't significantly faster than TMPGenc are simply performing flawed comparisons, period.

On a multi-processor system, yes CCE isn't much faster than TMPGenc. Other than that, it is between 2 and 4 times faster.

Remember the first pass in CCE is the creation of the .vaf file. So if you set it to 2-pass VBR you are actually doing 3-pass VBR. So a "2-pass" encode in CCE compared to a "2-pass" encode in TMPGenc is NOT a proper test. Learn how each encoder works, then make a comparison.

I think by far the biggest flaw in people's tests are that they fail to consider the fact that they just have a bottleneck which exists outside of the encoder. Unless you are using a raw avi souce, the speed at which you encode can only be as fast as the speed at which you decode your source. Even on a slow pc CCE can achieve double time per pass, but some codecs cannot even be decoded that fast. The same effect can be achieved by using certain filters or excessive filters. The limiting factor here is the speed at which the data is being fed into the encoder, NOT the speed of the encoder. With full D1 material I get double time encoding in CCE and about .8 in TMPGenc, but if I run a simple inverse telecine script I get about .65 in both encoders. Obviously this encode tells you nothing about the speed of the encoders, only of the speed of the filter.

Also don't forget that one of CCE's greatest advantages is that it can accept YUV sources natively whereas most other encoders have to convert it to RGB just to convert it back to YUV again. Given a YUV source and proper handling of it, ie: avisynth, this increases encoding speed by up to a whopping 40%.

The fact that CCE is significantly faster (2-4x) than TMPGenc has just been tested and retested on so many different levels than it amazes me some people still don't believe it. The author of TMPGenc himself has admitted it several times in interviews. This is not up for debate and its certainly not a coin toss. I love TMPGenc and it is a wonderful encoder but raw encoding speed just isn't one of its selling points.

adam has done his homework

[cooljoe666]

Well, there will always be newbies meaning they will not know what to search for. I have 2 brothers who ask me the dumbest questions....but theyre not really dumb questions in their mind.

There is no doubt when you weigh the pros and cons between CCE and TmpgEnc......CCE is clearly the winner. Quality and speed especially. You have to learn alittle with any program you use and you have to learn alittle bit more w/ CCE but the results are consistent unlike TMPGENC. I finally got tired of messing with the program so I switched to CCE using nothing more than CCE, Avisynth, FitCD, BBmpeg, and Virtualdub. You can use the TmpGenc (free) to extract the audio to MP2. Thanks to Fitcd there is no need to learn AviScript. I have converted several videos since switching and it's simplicity.

There is a good guide that will makes these things clear.....it tells how to convert to DVD but all you have to do is switch the settings to SVCD if that's what your after.............. http://www.donstevenson.net/myguides/myguides.html

[adam]

Also, I would really like to read some of the interviews with Mr. Hiroyuki Hori (is that the author of TMPGenc?)- that would be very interesting. Also, Adam, if you could point me to some of the tests indicating its speed, I would be most appreciative.

Those interviews with Hiroyuki Hori are old. I don't know if the there are any live links to them anymore. The only one I know of is here, http://www.cdrinfo.com/Sections/Articles/Specific.asp?ArticleHeadline=Cebit+2002+Show&index=7 but I think its an abridged version. He talks a little about how CCE is TMPGenc's main competitor and about how TMPGenc has always prioritized quality over speed, but in the full interview he mentions that he eventually wants to completely rewrite TMPGenc in a new programming language because using the current one, it will never be as fast as CCE.

There are just dozens of threads on this site and on www.doom9.net comparing CCE's speed to TMPGenc. All I can tell you is to do a forum search, you really shouldn't have any problems finding the data. I can't count the number of controlled tests that have been performed by a large group all reaching the same results.

http://www.tecoltd.com/enctest/enctest.htm

A long time ago Teco did a very extensive comparison of all the major commercial encoders and TMPGenc basically bombed in the speed department. TMPGenc has gotten substantially faster since then, and later versions of CCE have actually gotten slower since then, but honestly the margin really hasn't decreased all that much. CCE and now Mainconcept just blow away most of their competitors in the speed department.

I thought of another flaw that people may have in their comparisons between these two encoders. You've got to use a long clip to test with. 5 or 10 mins proves nothing, the whole point is actually make the encoder fully use the algorithms that it would have to in a full length encode. I'd say you'd need about a 30 min sample to do a proper speed test. Also, I've noticed that CCE doesn't reach maximum speed until about 10 mins into the encode.

To those people who mentioned Procoder. At normal settings Procoder is pretty fast and can compete in the quality department with other encoders marginally well, especially with interlaced material, but unlike most other encoders the higher quality level settings REALLY do make a difference. The difference between normal and high quality is pretty easily seen, and the same is true with the mastering level quality, and at these settings Procoder is easily the slowest encoder on the planet. In CCE I can encode an hour of D1 footage in about 30 mins. That same footage in Procoder's mastering quality would take about 30 hours. These speeds at these settings are pretty ridiculous but I personally don't care too much for Procoder's quality unless you use them, so I kinda have a hard time considering Procoder fast, but that's for my own purposes.

[andie41]

Thanks Adam, for your reply; however, as you mentioned the tests done by Teco are old (I might add, by computer standards, ancient), the CCE version was the professional version, and TMPGenc version 2.02.

The interview with Hiroyuki Hori was also quite old and it must have been. as you mentioned, an abridged version, because he certainly didn't talk about a big speed difference between TMPGenc and CCE. All he said was TMPGenc wasn't as fast as it should be. Also you mention that "there are just dozens of threads on this site and on www.doom9.net comparing CCE's speed to TMPGenc." I realize there are numerous opinions and anecdotal comments, but I was hoping for something more akin to the Teco review, only updated.

I happened upon a well written message in a forum on another site earlier tonight, and the writer certainly appears to know his stuff, but left me totally confused. He was quite high on the Mainconcept encoder and when I previously tried it, it was easily the worst of my tests. (in terms of artifacts.) Also he was quite positive about Procoder and negative on Ligos. Isn't Ligos the encoder for Procoder?

You mention the necessity of using longer clips, and it occurs to me that all these conflicting opinions arise because we are all doing different things, which of course affects our results. I rarely have a clip over 20 minutes. I am not making Divx or VCD's or trying to squeeze an extra minute onto my DVD. I am taking DV from my Digital 8 to the computer and encoding VBR CQ (quality 100) at around 9 Mbits. I am quite satisfied with the results. And when viewed on a TV, I can barely tell any diffence in any of these encoders, and only when I look very carefully.

I had seen the Teco test before, but it was interesting to go back and look at again. It was interesting looking at the results in bitrate viewer. The "Q" level, even for CCE, was 8. I would be very disappointed if my mpegs had a "Q" that high.

Anyway, thanks very much for you comments and time.

[SatStorm]

There is a rumour out there, that Hiroyuki Hori finally decide to rewrite TMPGenc. Maybe that explains the delay on new versions....

TMPGenc 3.0 gonna be "faster" based on those japanese rumours

[andie41]

There is a rumour out there, that Hiroyuki Hori finally decide to rewrite TMPGenc

That pretty well sums it up. The internet thrives on rumor and speculation - and it makes for enjoyment and sometimes frustration - it is just human nature.

I would certainly expect that people reading through this thread and Adam's statement and liink to Hiroyuki Hori's interview, came away believing that Hiroyuki Hori said TMPGenc, in its current version, will never be as fast as CCE. He did not say that in the link - Adam said that. I am not saying he never said it, just that all I honestly have to back that up is Adam's assurance that he said it. Even accepting that he said it and that CCE is in fact faster, it doesn't mean he said CCE is 500 percent faster.

This discussion is ongoing. I hope no one is offended - I enjoy discussion.

[sacajaweeda]

I use both CCE & TMPGEnc and the speed & quality of CCE makes me want to throw rocks at TMPGEnc. CCE gets all the encode work and all I use TMPGEnc for anymore really is multiplexing mpeg streams I've encoded with other apps, or sometimes encoding audio for MPEG2. I don't waste my time encoding video for DVD with it anymore. VCD & SVCD here and there maybe, but even that is getting rare.

[SatStorm]

@sacajaweeda: Determine better please: CCE with Avisynth vs TMPGenc plus as a standalone application is simply not fair to compare!
I don't like misunderstantings!